Do reviewer scores correlate with price?

After all, you would expect more expensive whiskies to be favoured by reviewers. And any discrepancy where a low cost whisky is favoured over a higher priced ones can be a great opportunity to identify “bargains”. This is known as arbitrage in the liquor industry, and is commonly used by wine enthusiasts.

So what do we see if we look at all whiskies in the dataset (which had a max of ~$3000)? For the sake of resolution, I’m going to show the results only for the <$400 whiskies.

There is not as much of an overall correlation as you might expect here (and if there is, it is more non-linear across this scale range). But the reason for this is that there are actually different discrete relationships depending on the price range you are looking at. For example, almost all whiskies over ~$150 a bottle get a decent score (i.e., above average).

To better illustrate that point, let’s restrict ourselves to more affordable <$100 whiskies.

OK, that is more what I expected to see. Yes, there is a positive correlation – but it is not overly strong. There are plenty of examples of low cost whiskies that score higher than more expensive offerings. So if you are looking to isolate “good buys”, it seems like this meta-critic score should provide you with a lot of options to consider.

Note that if you continue to subdivide the data into increasingly narrow price bands, the correlation may disappear again. It is really only across a reasonably wide range where the (admittedly weak) correlation matters at all.

In any case, since you aren’t likely to be trying to compare a $20 whisky with a $2000 one, I think we can safely discount any significant role of price on whisky quality within a reasonable price band. Overall, price more likely reflects relative rarity/scarcity, production complexity, or some perceptual “luxury” cachet.

A more important correlation to now consider is score by flavour profile.